AIoT人工智能物联网之deepstream

 import gi

    
 gi.require_version('Gst','1.0')
    
 gi.require_version('GstRtspServer','1.0')
    
 from gi.repository import Gst, GObject, GstRtspServer
    
  
    
 Gst.init(None)
    
 mainloop = GObject.MainLoop()          # 产生循环
    
 server = GstRtspServer.RTSPServer()    # 服务
    
 factory = GstRtspServer.RTSPMediaFactory() # 工程产生管道
    
 factory.set_launch(
    
     ' ( videotestsrc is-live=1 '
    
     '! video/x-raw,format=(string)NV12,width=(int)320,height=(int)180,framerate=20/1 '
    
     '! nvvidconv ! nvv4l2h264enc ! h264parse ! rtph264pay name=pay0 bitrate=40000 pt=96 ) '
    
 )
    
 # allow multiple connections
    
 factory.set_shared(True)     # 设置共享，可以多个连接
    
 mounts = server.get_mount_points()  # 产生挂载点
    
 mounts.add_factory('/live', factory) # 工程挂载上去
    
 print('rtsp://192.168.2.9:8554/live')
    
 server.attach(None)
    
 mainloop.run()
    
    
    
    
    AI助手

 import gi

    
 gi.require_version('Gst','1.0')
    
 gi.require_version('GstRtspServer','1.0')
    
 from gi.repository import Gst, GObject, GstRtspServer
    
  
    
 Gst.init(None)
    
 mainloop = GObject.MainLoop()
    
 server = GstRtspServer.RTSPServer()
    
 factory = GstRtspServer.RTSPMediaFactory()
    
 factory.set_launch(
    
     ' ( nvarguscamerasrc '  # 内存运到显存
    
     '! video/x-raw(memory:NVMM),width=(int)1280,height=(int)720,format=(string)NV12,framerate=(fraction)30/1 '
    
     '! nvv4l2h264enc ! h264parse ! rtph264pay name=pay0 bitrate=40000 pt=96 ) '
    
 )
    
  
    
 # allow multiple connections
    
 factory.set_shared(True)
    
 mounts = server.get_mount_points()
    
 mounts.add_factory('/live', factory)
    
 print('rtsp://192.168.2.9:8554/live')
    
 server.attach(None)
    
 mainloop.run()
    
    
    
    
    AI助手

 gst-launch-1.0 udpsrc port=5004 name=pay0 buffer-size=524288

    
 caps="application/x-rtp, media=video, clock-rate=90000,
    
 encoding-name=(string)H264, payload=96 " ! rtph264depay !
    
 h264parse ! nvh264dec ! videoconvert ! autovideosink -e
    
    
    
    
    AI助手

 gst-launch-1.0 -v nvarguscamerasrc bufapi-version=1 ! 'video/xraw(memory:NVMM),width=(int)1280,height=(int)720,format=(stri

    
 ng)NV12,framerate=(fraction)30/1' ! nvv4l2h264enc presetlevel=1 insert-sps-pps=1 bufapi-version=1 ! h264parse !
    
 rtph264pay pt=96 ! udpsink host=192.168.2.7 port=5004
    
 async=False sync=1
    
    
    
    
    AI助手

 在宿主机上执行，避免无法获取到图形界面的一些问题，

    
 715a93f1d755是ImageId 改为你自己的，注意GStreamer目录映
    
 射
    
 docker run -it --rm --net=host --runtime nvidia -e
    
 DISPLAY=$DISPLAY -v /tmp/argus_socket:/tmp/argus_socket -- device /dev/video0 -v
    
 ~/GStreamer:/opt/nvidia/deepstream/deepstream- 6.0/sources/deepstream_python_apps/apps/GStreamer
    
 715a93f1d755
    
    
    
    
    AI助手

 import sys

    
 sys.path.append('../')
    
 import gi
    
 gi.require_version('Gst', '1.0')
    
 from gi.repository import GObject, Gst
    
 from common.bus_call import bus_call
    
  
    
 # 本程序等价于 下面的脚本, 自动使用时注意把192.168.2.7改为自己的客户端IP
    
 # gst-launch-1.0 -v nvarguscamerasrc
    
 # ! 'video/x-raw(memory:NVMM),width=(int)1280,height=(int)720,format=(string)NV12,framerate=(fraction)30/1'
    
 # ! nvv4l2h264enc ! h264parse ! rtph264pay pt=96
    
 # ! udpsink host=192.168.2.7 port=5004
    
  
    
  
    
 Gst.init(None)
    
 pipeline = Gst.Pipeline()
    
 camerasrc  = Gst.ElementFactory.make("nvarguscamerasrc", "nv-arguscamerasrc") # 创建element
    
 caps = Gst.ElementFactory.make("capsfilter", "filter") # 格式
    
 caps.set_property("caps", Gst.Caps.from_string("video/x-raw(memory:NVMM),width=(int)1280,height=(int)720,format=(string)NV12,framerate=(fraction)30/1"))
    
 encoder = Gst.ElementFactory.make("nvv4l2h264enc", "encoder") # 创建编码器
    
 h264parse = Gst.ElementFactory.make("h264parse", "h264parse")
    
 rtppay = Gst.ElementFactory.make("rtph264pay", "rtppay")
    
 sink = Gst.ElementFactory.make("udpsink", "udpsink")
    
 camerasrc.set_property('bufapi-version', 1)
    
 encoder.set_property('preset-level', 1)
    
 encoder.set_property('insert-sps-pps', 1)
    
 encoder.set_property('bufapi-version', 1)
    
 rtppay.set_property("pt", 96)
    
 sink.set_property('host', '192.168.2.7')
    
 sink.set_property('port', 5004)
    
 sink.set_property('async', False)
    
 sink.set_property('sync', 1)
    
  
    
 pipeline.add(camerasrc ) # 加入到管道中
    
 pipeline.add(caps)
    
 pipeline.add(encoder)
    
 pipeline.add(h264parse)
    
 pipeline.add(rtppay)
    
 pipeline.add(sink)
    
  
    
 camerasrc.link(caps)  # 连接起来
    
 caps.link(encoder)
    
 encoder.link(h264parse)
    
 h264parse.link(rtppay)
    
 rtppay.link(sink)
    
  
    
 loop = GObject.MainLoop() # 循环
    
 bus = pipeline.get_bus()  # 总线
    
 bus.add_signal_watch()    # 信号连接器
    
 bus.connect("message", bus_call, loop)
    
 pipeline.set_state(Gst.State.PLAYING)
    
 try:
    
     loop.run()
    
 except:
    
     pass
    
 pipeline.set_state(Gst.State.NULL)
    
  
    
 # 摄像头RTP客户端Windows
    
 # gst-launch-1.0 udpsrc port=5004 name=pay0 buffer-size=524288 caps="application/x-rtp, media=video, clock-rate=90000, encoding-name=(string)H264, payload=96 " ! rtph264depay ! h264parse ! nvh264dec ! videoconvert ! autovideosink -e
    
  
    
    
    
    
    AI助手

 import sys

    
 sys.path.append('../')
    
 import gi
    
 gi.require_version('Gst', '1.0')
    
 gi.require_version('GstRtspServer', '1.0')  # 注册一个rtsp server
    
 from gi.repository import GObject, Gst, GstRtspServer
    
 from common.bus_call import bus_call
    
  
    
 GObject.threads_init()
    
 Gst.init(None)
    
  
    
 pipeline = Gst.Pipeline()
    
 camerasrc  = Gst.ElementFactory.make("nvarguscamerasrc", "nv-arguscamerasrc")
    
 caps = Gst.ElementFactory.make("capsfilter", "filter")
    
 caps.set_property("caps", Gst.Caps.from_string("video/x-raw(memory:NVMM),width=(int)1280,height=(int)720,format=(string)NV12,framerate=(fraction)30/1"))
    
 encoder = Gst.ElementFactory.make("nvv4l2h264enc", "encoder")
    
 h264parse = Gst.ElementFactory.make("h264parse", "h264parse")
    
 rtppay = Gst.ElementFactory.make("rtph264pay", "rtppay")
    
 sink = Gst.ElementFactory.make("udpsink", "udpsink")
    
 camerasrc.set_property('bufapi-version', 1)
    
 encoder.set_property('preset-level', 1)
    
 encoder.set_property('insert-sps-pps', 1)
    
 encoder.set_property('bufapi-version', 1)
    
 rtppay.set_property("pt", 96)
    
 sink.set_property('host', '192.168.2.9') # 本机IP
    
 sink.set_property('port', 5004)
    
 sink.set_property('async', False)
    
 sink.set_property('sync', 1)
    
  
    
 pipeline.add(camerasrc )
    
 pipeline.add(caps)
    
 pipeline.add(encoder)
    
 pipeline.add(h264parse)
    
 pipeline.add(rtppay)
    
 pipeline.add(sink)
    
  
    
 camerasrc .link(caps)
    
 caps.link(encoder)
    
 encoder.link(h264parse)
    
 h264parse.link(rtppay)
    
 rtppay.link(sink)
    
  
    
 server = GstRtspServer.RTSPServer.new()  # 产生一个设置实例
    
 server.props.service = "%d" % 8554   # 整型，设置端口号
    
 server.attach(None)
    
 factory = GstRtspServer.RTSPMediaFactory.new() # 产生工厂,提供多媒体数据
    
 factory.set_launch(
    
     "( udpsrc name=pay0 port=%d buffer-size=524288 caps=\"application/x-rtp, media=video, clock-rate=90000, encoding-name=(string)H264, payload=96 \" )" % (
    
     5004))
    
 factory.set_shared(True) # 共享
    
 server.get_mount_points().add_factory("/live", factory) # 挂载
    
  
    
 loop = GObject.MainLoop()
    
 bus = pipeline.get_bus()
    
 bus.add_signal_watch()
    
 bus.connect("message", bus_call, loop)
    
 pipeline.set_state(Gst.State.PLAYING)
    
 print("rtsp://192.168.2.9:8554/live")
    
 try:
    
     loop.run()
    
 except:
    
     pass
    
 pipeline.set_state(Gst.State.NULL)
    
    
    
    
    AI助手

 import sys

    
 sys.path.append('../')
    
  
    
 import gi
    
 gi.require_version('Gst', '1.0')
    
 gi.require_version('GstRtspServer', '1.0')
    
  
    
 from gi.repository import GObject, Gst, GstRtspServer
    
 from common.bus_call import bus_call
    
  
    
 GObject.threads_init()
    
 Gst.init(None)
    
  
    
 pipeline = Gst.Pipeline()
    
 camerasrc = Gst.ElementFactory.make("nvarguscamerasrc", "nv-arguscamerasrc")
    
 caps = Gst.ElementFactory.make("capsfilter", "filter")
    
 caps.set_property("caps", Gst.Caps.from_string("video/x-raw(memory:NVMM),width=(int)1280,height=(int)720,format=(string)NV12,framerate=(fraction)30/1"))
    
  
    
 # Create nvstreammux instance to form batches from one or more sources.
    
 streammux = Gst.ElementFactory.make("nvstreammux", "Stream-muxer")
    
 streammux.set_property('width', 1280)
    
 streammux.set_property('height', 720)
    
 streammux.set_property('batch-size', 1)
    
 streammux.set_property('batched-push-timeout', 4000000)
    
 # Use nvinfer to run inferencing on decoder's output,
    
 # behaviour of inferencing is set through config file
    
 pgie = Gst.ElementFactory.make("nvinfer", "primary-inference")       # 主推理
    
 pgie.set_property('config-file-path', "pgie_yolov4_tiny_tao_config.txt") # 设置配置文件
    
 # Use convertor to convert from NV12 to RGBA as required by nvosd
    
 nvvidconv = Gst.ElementFactory.make("nvvideoconvert", "convertor")   # 转换
    
 # Create OSD to draw on the converted RGBA buffer
    
 nvosd = Gst.ElementFactory.make("nvdsosd", "onscreendisplay")   # 元数据写入图片
    
 nvvidconv_postosd = Gst.ElementFactory.make("nvvideoconvert", "convertor_postosd") # 转换回来
    
 # Create a caps filter
    
 caps_postosd = Gst.ElementFactory.make("capsfilter", "caps_postosd_filter")
    
 caps_postosd.set_property("caps", Gst.Caps.from_string("video/x-raw(memory:NVMM), format=I420"))
    
  
    
 encoder = Gst.ElementFactory.make("nvv4l2h264enc", "encoder")
    
 h264parse = Gst.ElementFactory.make("h264parse", "h264parse")
    
 rtppay = Gst.ElementFactory.make("rtph264pay", "rtppay")
    
 sink = Gst.ElementFactory.make("udpsink", "udpsink")
    
 camerasrc.set_property('bufapi-version', 1)
    
 encoder.set_property('preset-level', 1)
    
 encoder.set_property('insert-sps-pps', 1)
    
 encoder.set_property('bufapi-version', 1)
    
 rtppay.set_property("pt", 96)
    
 sink.set_property('host', '192.168.2.9') # 本机IP
    
 sink.set_property('port', 5004)
    
 sink.set_property('async', False)
    
 sink.set_property('sync', 1)
    
  
    
 pipeline.add(camerasrc)
    
 pipeline.add(caps)
    
  
    
 # 新增
    
 pipeline.add(streammux)
    
 pipeline.add(pgie)
    
 pipeline.add(nvvidconv)
    
 pipeline.add(nvosd)
    
 pipeline.add(nvvidconv_postosd)
    
 pipeline.add(caps_postosd)
    
  
    
 pipeline.add(encoder)
    
 pipeline.add(h264parse)
    
 pipeline.add(rtppay)
    
 pipeline.add(sink)
    
  
    
 sinkpad = streammux.get_request_pad("sink_0") # 可以有多个数据源，选择一个数据源
    
 srcpad = caps.get_static_pad("src")           # 获取src源
    
  
    
 camerasrc .link(caps)
    
 srcpad.link(sinkpad)
    
 streammux.link(pgie)
    
  
    
 pgie.link(nvvidconv)
    
 nvvidconv.link(nvosd)
    
 nvosd.link(nvvidconv_postosd)
    
 nvvidconv_postosd.link(caps_postosd)
    
  
    
 caps_postosd.link(encoder)
    
 encoder.link(h264parse)
    
 h264parse.link(rtppay)
    
 rtppay.link(sink)
    
  
    
 server = GstRtspServer.RTSPServer.new()
    
 server.props.service = "%d" % 8554
    
 server.attach(None)
    
 factory = GstRtspServer.RTSPMediaFactory.new()
    
 factory.set_launch(
    
     "( udpsrc name=pay0 port=%d buffer-size=524288 caps=\"application/x-rtp, media=video, clock-rate=90000, encoding-name=(string)H264, payload=96 \" )" % (
    
     5004))
    
 factory.set_shared(True)
    
 server.get_mount_points().add_factory("/live", factory)
    
  
    
 loop = GObject.MainLoop()
    
 bus = pipeline.get_bus()
    
 bus.add_signal_watch()
    
 bus.connect("message", bus_call, loop)
    
 pipeline.set_state(Gst.State.PLAYING)
    
 print("rtsp://192.168.2.9:8554/live")
    
 try:
    
     loop.run()
    
 except:
    
     pass
    
 pipeline.set_state(Gst.State.NULL)
    
    
    
    
    AI助手

 # Following properties are mandatory when engine files are not specified:

    
 #   int8-calib-file(Only in INT8)
    
 #   Caffemodel mandatory properties: model-file, proto-file, output-blob-names
    
 #
    
 # Mandatory properties for detectors:
    
 #   num-detected-classes
    
 #
    
 # Following properties are always recommended:
    
 #   batch-size(Default=1)
    
 #
    
 # Other optional properties:
    
 #   net-scale-factor(Default=1), network-mode(Default=0 i.e FP32),
    
 #   model-color-format(Default=0 i.e. RGB) model-engine-file, labelfile-path,
    
 #   mean-file, gie-unique-id(Default=0), offsets, process-mode (Default=1 i.e. primary),
    
 #   custom-lib-path, network-mode(Default=0 i.e FP32)
    
 #
    
 # The values in the config file are overridden by values set through GObject
    
 # properties.
    
  
    
 # Gst-nvinfer 插件根据网络要求对输入帧执行转换（格式转换和缩放），并将转换后的数据传递给低级库。
    
 # 低级库对转换后的帧进行预处理（执行归一化和均值减法）并生成最终的浮点 RGB/BGR/GRAY 平面数据，该数据传递给 TensorRT 引擎进行推理。低级库生成的输出类型取决于网络类型
    
 # y=net-scale-factor*(x-mean)
    
  
    
 [property]
    
 # 多个GPU的情况下，元件所使用的GPU设备ID
    
 gpu-id=0
    
 # 低级库对转换后的帧进行预处理（执行归一化和均值减法）并生成最终的浮点 RGB/BGR/GRAY 平面数据，该数据传递给 TensorRT 引擎进行推理。低级库生成的输出类型取决于网络类型
    
 # y=net-scale-factor*(x-mean)
    
 net-scale-factor=0.0039215697906911373
    
 # caffe 模型文件路径
    
 model-file=../../../../samples/models/Primary_Detector/resnet10.caffemodel
    
 proto-file=../../../../samples/models/Primary_Detector/resnet10.prototxt
    
 # 模型预先生成的序列化引擎文件的绝对路径，如果有这个就不在看 model文件
    
 model-engine-file=../../../../samples/models/Primary_Detector/resnet10.caffemodel_b1_gpu0_int8.engine
    
 labelfile-path=../../../../samples/models/Primary_Detector/labels.txt
    
 # 当模型降低到8位时，验证的库
    
 int8-calib-file=../../../../samples/models/Primary_Detector/cal_trt.bin
    
 force-implicit-batch-dim=1
    
 batch-size=1
    
  
    
 # Integer
    
 # 0: FP32
    
 # 1: INT8
    
 # 2: FP16
    
 network-mode=1
    
 num-detected-classes=4  # 目标检测类别
    
 # 推理时要跳过的连续批数
    
 interval=0
    
 # 要分配给 nvinfer 实例的唯一组件 ID。用于标识由实例生成的元数据
    
 gie-unique-id=1
    
 # 输出检测内容的元数据名字
    
 output-blob-names=conv2d_bbox;conv2d_cov/Sigmoid
    
  
    
 [class-attrs-all]
    
 # 目标检测阈值
    
 pre-cluster-threshold=0.2
    
 # DBSCAN algorithm 聚和算法参数
    
 eps=0.2
    
 # 矩形合并的阈值
    
 group-threshold=1
    
    
    
    
    AI助手

 export CUDA_HOME=/usr/local/cuda

    
 export PATH=/usr/local/cuda/bin:$PATH
    
 export LD_LIBRARY_PATH=/usr/local/cuda/lib64/:$LD_LIBRARY_PATH
    
    
    
    
    AI助手

 sudo mv /usr/lib/aarch64-linux-gnu/libnvinfer_plugin.so.8.x.y ${HOME}/libnvinfer_plugin.so.8.x.y.bak // backup original libnvinfer_plugin.so.x.y

    
 sudo cp $TRT_SOURCE/build/libnvinfer_plugin.so.8.m.n /usr/lib/aarch64-linux-gnu/libnvinfer_plugin.so.8.x.y
    
 sudo ldconfig
    
    
    
    
    AI助手

 docker run -it --rm --net=host --runtime nvidia -e

    
 DISPLAY=$DISPLAY -v /tmp/argus_socket:/tmp/argus_socket -- device /dev/video0 -v
    
 ~/GStreamer:/opt/nvidia/deepstream/deepstream- 6.0/sources/deepstream_python_apps/apps/GStreamer
    
 715a93f1d755
    
    
    
    
    AI助手

 net-scale-factor=1.0

    
 offsets=103.939;116.779;123.68
    
 infer-dims=3;384;1248
    
 tlt-model-key=NDZsYWZyODZpMWZmNXNkcHFpNnI3djBnNjk6M2ZkMWJlNGQtOTI1Yi00MWYzLTk2NDgtYWEwYzEwODFiNmJm
    
 network-type=0
    
 num-detected-classes=1
    
 model-color-format=1
    
 maintain-aspect-ratio=0
    
 output-tensor-meta=0
    
    
    
    
    AI助手

 [property]

    
 gpu-id=0
    
 net-scale-factor=1.0
    
 offsets=103.939;116.779;123.68
    
 model-color-format=1
    
 labelfile-path=./export/labels.txt
    
 model-engine-file=./export/trt.engine
    
 #int8-calib-file=../../models/yolov4-tiny/cal.bin
    
 tlt-encoded-model=./export/yolov4_cspdarknet_tiny_epoch_080.etlt
    
 tlt-model-key=NDZsYWZyODZpMWZmNXNkcHFpNnI3djBnNjk6M2ZkMWJlNGQtOTI1Yi00MWYzLTk2NDgtYWEwYzEwODFiNmJm
    
 infer-dims=3;384;1248
    
 maintain-aspect-ratio=0
    
 uff-input-order=0
    
 uff-input-blob-name=Input
    
 batch-size=1
    
 ## 0=FP32, 1=INT8, 2=FP16 mode
    
 network-mode=1
    
 num-detected-classes=1
    
 interval=0
    
 gie-unique-id=1
    
 is-classifier=0
    
 #network-type=0
    
 cluster-mode=3
    
 output-blob-names=BatchedNMS
    
 parse-bbox-func-name=NvDsInferParseCustomBatchedNMSTLT
    
 custom-lib-path=./export/libnvds_infercustomparser_tao.so
    
  
    
 [class-attrs-all]
    
 pre-cluster-threshold=0.3
    
 roi-top-offset=0
    
 roi-bottom-offset=0
    
 detected-min-w=0
    
 detected-min-h=0
    
 detected-max-w=0
    
 detected-max-h=0
    
    
    
    
    AI助手

    rm -rf ~/.cache/gstreamer-1.0/
    
    AI助手

 cd /opt/nvidia/deepstream/deepstream-6.0/sources/deepstream_python_apps/a

    
 pps/GStreamer
    
    
    
    
    AI助手

 import sys

    
 sys.path.append('../')
    
  
    
 import gi
    
 gi.require_version('Gst', '1.0')
    
 gi.require_version('GstRtspServer', '1.0')
    
  
    
 from gi.repository import GObject, Gst, GstRtspServer
    
 from common.bus_call import bus_call
    
  
    
 GObject.threads_init()
    
 Gst.init(None)
    
  
    
 pipeline = Gst.Pipeline()
    
 camerasrc = Gst.ElementFactory.make("nvarguscamerasrc", "nv-arguscamerasrc")
    
 caps = Gst.ElementFactory.make("capsfilter", "filter")
    
 caps.set_property("caps", Gst.Caps.from_string("video/x-raw(memory:NVMM),width=(int)1280,height=(int)720,format=(string)NV12,framerate=(fraction)30/1"))
    
  
    
 # Create nvstreammux instance to form batches from one or more sources.
    
 streammux = Gst.ElementFactory.make("nvstreammux", "Stream-muxer")
    
 streammux.set_property('width', 1280)
    
 streammux.set_property('height', 720)
    
 streammux.set_property('batch-size', 1)
    
 streammux.set_property('batched-push-timeout', 4000000)
    
 # Use nvinfer to run inferencing on decoder's output,
    
 # behaviour of inferencing is set through config file
    
 pgie = Gst.ElementFactory.make("nvinfer", "primary-inference")       # 主推理
    
 pgie.set_property('config-file-path', "pgie_yolov4_tiny_tao_config.txt") # 设置配置文件
    
 # Use convertor to convert from NV12 to RGBA as required by nvosd
    
 nvvidconv = Gst.ElementFactory.make("nvvideoconvert", "convertor")   # 转换
    
 # Create OSD to draw on the converted RGBA buffer
    
 nvosd = Gst.ElementFactory.make("nvdsosd", "onscreendisplay")   # 元数据写入图片
    
 nvvidconv_postosd = Gst.ElementFactory.make("nvvideoconvert", "convertor_postosd") # 转换回来
    
 # Create a caps filter
    
 caps_postosd = Gst.ElementFactory.make("capsfilter", "caps_postosd_filter")
    
 caps_postosd.set_property("caps", Gst.Caps.from_string("video/x-raw(memory:NVMM), format=I420"))
    
  
    
 encoder = Gst.ElementFactory.make("nvv4l2h264enc", "encoder")
    
 h264parse = Gst.ElementFactory.make("h264parse", "h264parse")
    
 rtppay = Gst.ElementFactory.make("rtph264pay", "rtppay")
    
 sink = Gst.ElementFactory.make("udpsink", "udpsink")
    
 camerasrc.set_property('bufapi-version', 1)
    
 encoder.set_property('preset-level', 1)
    
 encoder.set_property('insert-sps-pps', 1)
    
 encoder.set_property('bufapi-version', 1)
    
 rtppay.set_property("pt", 96)
    
 sink.set_property('host', '192.168.2.9') # 本机IP
    
 sink.set_property('port', 5004)
    
 sink.set_property('async', False)
    
 sink.set_property('sync', 1)
    
  
    
 pipeline.add(camerasrc)
    
 pipeline.add(caps)
    
  
    
 # 新增
    
 pipeline.add(streammux)
    
 pipeline.add(pgie)
    
 pipeline.add(nvvidconv)
    
 pipeline.add(nvosd)
    
 pipeline.add(nvvidconv_postosd)
    
 pipeline.add(caps_postosd)
    
  
    
 pipeline.add(encoder)
    
 pipeline.add(h264parse)
    
 pipeline.add(rtppay)
    
 pipeline.add(sink)
    
  
    
 sinkpad = streammux.get_request_pad("sink_0") # 可以有多个数据源，选择一个数据源
    
 srcpad = caps.get_static_pad("src")           # 获取src源
    
  
    
 camerasrc .link(caps)
    
 srcpad.link(sinkpad)
    
 streammux.link(pgie)
    
  
    
 pgie.link(nvvidconv)
    
 nvvidconv.link(nvosd)
    
 nvosd.link(nvvidconv_postosd)
    
 nvvidconv_postosd.link(caps_postosd)
    
  
    
 caps_postosd.link(encoder)
    
 encoder.link(h264parse)
    
 h264parse.link(rtppay)
    
 rtppay.link(sink)
    
  
    
 server = GstRtspServer.RTSPServer.new()
    
 server.props.service = "%d" % 8554
    
 server.attach(None)
    
 factory = GstRtspServer.RTSPMediaFactory.new()
    
 factory.set_launch(
    
     "( udpsrc name=pay0 port=%d buffer-size=524288 caps=\"application/x-rtp, media=video, clock-rate=90000, encoding-name=(string)H264, payload=96 \" )" % (
    
     5004))
    
 factory.set_shared(True)
    
 server.get_mount_points().add_factory("/live", factory)
    
  
    
 loop = GObject.MainLoop()
    
 bus = pipeline.get_bus()
    
 bus.add_signal_watch()
    
 bus.connect("message", bus_call, loop)
    
 pipeline.set_state(Gst.State.PLAYING)
    
 print("rtsp://192.168.2.9:8554/live")
    
 try:
    
     loop.run()
    
 except:
    
     pass
    
 pipeline.set_state(Gst.State.NULL)
    
  
    
    
    
    
    AI助手